5 ms·

Relevant / obligatory xkcd: https://xkcd.com/2030/ https://xkcd.com/2030/ To quote Randall, "don't trust voting software and don't listen to anyone who tells y

Relevant / obligatory xkcd: https://xkcd.com/2030/ https://xkcd.com/2030/ To quote Randall, "don't trust voting software and don't listen to anyone who tells you it's safe."

throwaway894345 6y ago
I don't know about the current state of voting software, but space shuttle software has a pretty good track record. We know how to produce good software just like we know how to build good elevators and so on. The perceived problem that Randall is observing is that there's a lot of software out there for social media networks or news sites or games and all manner of other software which isn't life or death and for which faults are perfectly acceptable (at least to the extent that we don't bother to regulate those faults more aggressively). So what Randall (and many others) perceives as "we're bad at software" is actually "we're pretty good at software, it's just that most software doesn't need to be space-shuttle/voting-systems grade, and we don't waste money making it to that caliber". The most valid reason to be opposed to voting software (and it's my personal reason) is that it's almost certainly easier to compromise an electronic voting system--even a very high quality system--than it is to compromise a paper ballot election. With a paper ballot election, it's probably easier to just rig public opinion by compromising social media companies. I would also suspect that we could make it even harder to rig a paper ballot election if we elected by popular vote such that (for tight elections) the malicious actor would have to compromise a lot more than just a few key counties in a few swing states.
- lultimouomo 6y ago
  There is a fundamental difference between space shuttle and voting software: it is very rare for people directly involved in developing the first to have incentive to make a shuttle crash; it is not so clear that people involved in voting systems can't have a incentive to make elections rigged. The possibility of malicious code is much more of a problem when people stand to profit from a bug.
  - AnimalMuppet 6y ago
    People involved in voting systems still have an incentive to make it impossible for their opponents to rig an election. So get both sides involved...
    - lultimouomo 6y ago
      Get both sides involved could just mean having an auction... I'm not saying it's impossible to make a publicly verifyable, tamper proof system, but it's definitely harder than flying a shuttle.
      - AnimalMuppet 6y ago
        > Get both sides involved could just mean having an auction... No, because the loser of the auction would have every incentive to cry foul.
        lultimouomo 6y ago
        They wouldn't know they lost the auction until the result have come out; at that point crying foul would mean admitting they were involved in trying to steal an election; and since verifying the correctness of the voting system would be extremely hard they might not even be able to dispute the result.
      - tzs 6y ago
        It's actually quite a bit easier than flying a space shuttle. See [1], or for a less technical description [2]. [1] https://www.usenix.org/legacy/event/evt08/tech/full_papers/chaum/chaum_html/index.html https://www.usenix.org/legacy/event/evt08/tech/full_papers/c... [2] https://en.wikipedia.org/wiki/Scantegrity https://en.wikipedia.org/wiki/Scantegrity
  - tzs 6y ago
    > it is not so clear that people involved in voting systems can't have a incentive to make elections rigged It is not so clear that there is a reasonable mechanism by which they could rig elections. When they make the hardware and software they have no idea what candidates will actually be on the ballots when the machines are used.
- shakna 6y ago
  > I don't know about the current state of voting software, but space shuttle software has a pretty good track record. Voting software isn't that. The current state of voting software is best described as "horrifying". That's not to say it has been exploited, the only time that _may_ have happened _might_ be the Georgia, 2016, case, but as the evidence was destroyed, we have no answers. [1] Whilst this year's DefCon didn't go ahead, there are have been voting machines at several of the conferences in past years. Like last year, where vulnerabilities discovered in 2016 were still present. [0] The experts in the matter call for paper ballots. They are scalable, and far harder to compromise, by _several_ orders of magnitude. [0] https://www.washingtonpost.com/business/2019/08/12/def-con-hackers-lawmakers-came-together-tackle-holes-election-security/ https://www.washingtonpost.com/business/2019/08/12/def-con-h... [1] https://slate.com/technology/2017/10/georgia-destroyed-election-data-right-after-a-lawsuit-alleged-the-system-was-vulnerable.html https://slate.com/technology/2017/10/georgia-destroyed-elect...
  - throwaway894345 6y ago
    Note that I agree with the experts, but the point I'm raising is that we could have space-shuttle grade voting software--the problem isn't technical but rather social, organizational, and political. That said, even space-shuttle grade voting software is probably not more secure than paper ballots.
    - shakna 6y ago
      > the problem isn't technical but rather social, organizational, and political I would disagree with you here. The problem isn't social, or political. The problem is one of economics. It simply isn't in the interests of those manufacturing the voting machines to make them secure. They go unpunished for flaws, and it isn't legal, most of the time, for anyone to try and peak beneath the covers and see if there are flaws present. They have no incentive to make things secure, and can simply attack anyone who tries to look at their security, with a high degree of success. In a simple cost/benefit analysis, it is cheaper to ignore security. They will continue to get the government contracts, regardless of whether or not they are deemed to be more or less secure than their competition. Why would they bother?

Relevant / obligatory xkcd: https://xkcd.com/2030/ https://xkcd.com/2030/ To quote Randall, "don't trust voting software and don't listen to anyone who tells you it's safe."

throwaway894345 6y ago

I don't know about the current state of voting software, but space shuttle software has a pretty good track record. We know how to produce good software just like we know how to build good elevators and so on. The perceived problem that Randall is observing is that there's a lot of software out there for social media networks or news sites or games and all manner of other software which isn't life or death and for which faults are perfectly acceptable (at least to the extent that we don't bother to regulate those faults more aggressively). So what Randall (and many others) perceives as "we're bad at software" is actually "we're pretty good at software, it's just that most software doesn't need to be space-shuttle/voting-systems grade, and we don't waste money making it to that caliber". The most valid reason to be opposed to voting software (and it's my personal reason) is that it's almost certainly easier to compromise an electronic voting system--even a very high quality system--than it is to compromise a paper ballot election. With a paper ballot election, it's probably easier to just rig public opinion by compromising social media companies. I would also suspect that we could make it even harder to rig a paper ballot election if we elected by popular vote such that (for tight elections) the malicious actor would have to compromise a lot more than just a few key counties in a few swing states.

lultimouomo 6y ago
There is a fundamental difference between space shuttle and voting software: it is very rare for people directly involved in developing the first to have incentive to make a shuttle crash; it is not so clear that people involved in voting systems can't have a incentive to make elections rigged. The possibility of malicious code is much more of a problem when people stand to profit from a bug.
- AnimalMuppet 6y ago
  People involved in voting systems still have an incentive to make it impossible for their opponents to rig an election. So get both sides involved...
  - lultimouomo 6y ago
    Get both sides involved could just mean having an auction... I'm not saying it's impossible to make a publicly verifyable, tamper proof system, but it's definitely harder than flying a shuttle.
    - AnimalMuppet 6y ago
      > Get both sides involved could just mean having an auction... No, because the loser of the auction would have every incentive to cry foul.
      - lultimouomo 6y ago
        They wouldn't know they lost the auction until the result have come out; at that point crying foul would mean admitting they were involved in trying to steal an election; and since verifying the correctness of the voting system would be extremely hard they might not even be able to dispute the result.
    - tzs 6y ago
      It's actually quite a bit easier than flying a space shuttle. See [1], or for a less technical description [2]. [1] https://www.usenix.org/legacy/event/evt08/tech/full_papers/chaum/chaum_html/index.html https://www.usenix.org/legacy/event/evt08/tech/full_papers/c... [2] https://en.wikipedia.org/wiki/Scantegrity https://en.wikipedia.org/wiki/Scantegrity
- tzs 6y ago
  > it is not so clear that people involved in voting systems can't have a incentive to make elections rigged It is not so clear that there is a reasonable mechanism by which they could rig elections. When they make the hardware and software they have no idea what candidates will actually be on the ballots when the machines are used.
shakna 6y ago
> I don't know about the current state of voting software, but space shuttle software has a pretty good track record. Voting software isn't that. The current state of voting software is best described as "horrifying". That's not to say it has been exploited, the only time that _may_ have happened _might_ be the Georgia, 2016, case, but as the evidence was destroyed, we have no answers. [1] Whilst this year's DefCon didn't go ahead, there are have been voting machines at several of the conferences in past years. Like last year, where vulnerabilities discovered in 2016 were still present. [0] The experts in the matter call for paper ballots. They are scalable, and far harder to compromise, by _several_ orders of magnitude. [0] https://www.washingtonpost.com/business/2019/08/12/def-con-hackers-lawmakers-came-together-tackle-holes-election-security/ https://www.washingtonpost.com/business/2019/08/12/def-con-h... [1] https://slate.com/technology/2017/10/georgia-destroyed-election-data-right-after-a-lawsuit-alleged-the-system-was-vulnerable.html https://slate.com/technology/2017/10/georgia-destroyed-elect...
- throwaway894345 6y ago
  Note that I agree with the experts, but the point I'm raising is that we could have space-shuttle grade voting software--the problem isn't technical but rather social, organizational, and political. That said, even space-shuttle grade voting software is probably not more secure than paper ballots.
  - shakna 6y ago
    > the problem isn't technical but rather social, organizational, and political I would disagree with you here. The problem isn't social, or political. The problem is one of economics. It simply isn't in the interests of those manufacturing the voting machines to make them secure. They go unpunished for flaws, and it isn't legal, most of the time, for anyone to try and peak beneath the covers and see if there are flaws present. They have no incentive to make things secure, and can simply attack anyone who tries to look at their security, with a high degree of success. In a simple cost/benefit analysis, it is cheaper to ignore security. They will continue to get the government contracts, regardless of whether or not they are deemed to be more or less secure than their competition. Why would they bother?